But software is, what, less than a century old? What possible evolutionary differences could have risen in that time frame?
I mean, the number zero is very unintuitive and the ancient Romans/greeks/chinese/whatever didn't discover it, despite the fact all evidence shows they were very smart in many things.
Yet you want me to believe that men understand that arrays are zero based better than women.
Apply that to any software principle.
But guns are, what, less than a millennia old? What possible evolutionary differences could have arisen in that time frame? No way that women and men would perform differently in competitive shooting [0], right? And as it turns out, men shoot faster, women shoot more accurately.
There's all sorts of things that have only existed for a century or two where men and women perform differently. Adaptation for one task often has effects on similar tasks, after all.
This doesn't mean men are better than women at CS! But "this field didn't exist a century ago" isn't sufficient reason to assert no differences can exist.
0: http://www.realbiathlon.com/2011/11/who-shoots-better-men-or...
I think there are real differences, other than the two I mentioned, but there is probably a lot of overlaps where you can find plenty of exceptions.